Innovation and Ontologies

186 Classification – The OntoCube
Other examples for general ontologies are the large general-purpose, formal Standard Upper
Ontology 220 (SUO)(Pease & Niles, 2002) and Sowa’s top-level ontology which has been derived from a
variety of sources in logic, linguistics, philosophy and artificial intelligence (Sowa, 2002).
5.2.5 Representation Ontologies
Representation ontologies (van Heijst, Schreiber & Wielinga, 1997) capture the representation
primitives used to formalize knowledge. They provide representational entities (i.e. frames)
without stating what is represented and how representation should take place (Batemann, 1991;
Fensel, 2001; Gruber, 1993; Mizoguchi, 2004a; Pirlein, 1995).
The most renowned example for a representation ontology is the Frame Ontology 221 (Gruber,
1993). This ontology captures conventions used in object-centered knowledge representation
systems, i.e. classes, subclasses, attributes, values, relations and axioms (Fensel, 2001; Benjamins
& Gómez-Pérez, in prep.). Due to its translator functionalities, the Frame Ontology enables
sharing of ontologies by use of different representation systems (Gómez-Pérez, Fernández-López
& Corcho, 2002).
5.3 Formality
[…]an ontology varies […] in the degree of formality […].
Formality […] can vary from natural language
to formal logic. (Jasper & Uschold, 1999)
According to the classification of (Uschold, 1996; Uschold & Grueninger, 1996), the dimension
of formality comprises four degrees of abstraction, by which an ontology can be realized. With
respect to the degree of formality needed for an application at hand, Gandon states in 2002:
“The final formal degree of the ontology depends on the use intended […]. [T]he
formalization task does not consist of replacing an informal version by a formal one, but to
augment an informal version with the relevant formal aspect needed by the operational
system. The purpose […] is to develop the formal counterpart of interesting and relevant
semantic aspects of the informal ontology in order to obtain a documented […] operational
ontology […]. [T]he ontologist will stop his progression on the continuum between informal
and formal ontology as soon as he has reached the formal level necessary and sufficient for
his system.”(Gandon, 2002b)
The degree of formality is correlated positively to the degree of machine processability and
correlated negatively with the usefulness for communication (cf. figure 64). It follows that a
natural language ontology is well suited to support communication in cases where automation is
of little importance (Fox & Grueninger, 1998; Fernández-López, Gómez-Pérez & Juristo, 1997).
220
More information can be found on the homepage of the Standard Upper Ontology Working Group of IEEE, at http://suo.ieee.org/
(2008/04/28).
221
More information is available online at: http://www-ksl.stanford.edu/people/brauch/demo/frame-ontology/ (2007/12/28).